Two silver wedding rings lie next to each other on a wooden table.
(Annette Riedl/picture alliance via Getty Images)
Reading Time: 7 minutes

While race intersects with every aspect of American life, hard data on that is limited when it comes to taxation. The IRS does not collect taxpayers’ racial or ethnic identity.

New research using novel methodology — starting with Survey of Consumer Finances household survey data, creating tax units and running a tax calculator against it — supports what Black legal scholars have long posited: that racial disparities are baked into the tax code. 

The analysis, from the Urban-Brookings Tax Policy Center, looked at the impact of the “marriage penalty,” when a couple pays more in federal income tax than the combined bill if they were single. That generally occurs when each spouse earns around the same amount of income. 

In situations where a couple pays less, they’re receiving a marriage bonus. 

The new analysis shows that Black couples are more likely to experience marriage penalties than white couples, even controlling for income. Under the 2017 Tax Cuts and Jobs Act that’s still in place, 46% of Black couples paid marriage penalties, compared to 43% of white couples. Marriage bonuses are more stark: only 36% of Black couples get them, compared to 43% of white couples.

The Tax Policy Center’s paper, published last month, builds on work by Dorothy Brown, a Georgetown University law professor who wrote The Whiteness of Wealth. Black women have worked outside the home more than white wives throughout history, she noted, whereas the marriage bonus often goes to couples with one spouse who stays home. The new paper attributed the presence of children and differences in income distribution as additional reasons for the disparities. 

The Center for Public Integrity sat down with Janet Holtzblatt, senior fellow at the Tax Policy Center and the study’s co-author, to learn more about racial disparities in the joint tax return and how agencies are working to make it easier to study the impact. 

This conversation has been edited for length and clarity.

Q: What findings surprised you while working on this paper?

To some extent our findings confirmed what others have been saying. Going back to the 1990s, there were several Black legal scholars who had looked at the data and indicated that it was likely that Black married couples would incur larger penalties than white couples. And the premise was that Black couples were more likely to have two spouses who were working and for those two spouses to have equal earnings. And our findings confirmed that. 

I expected that would be the case in the 1990s, when the scholars were first making this analysis and conclusions. And the reason I thought it’d be likely to happen in the 1990s was that … the tax rate structure and standard deduction was structured in a way that gave rise to marriage penalties. When two people got married and they were both working, their combined income as a married couple would push them into a higher tax rate. In the ‘90s, amongst all couples, you would expect a higher probability of marriage penalties. 

The other aspect was that between the 1970s and the 1990s, there were more and more two-earner couples in the labor market. I anticipated that would continue to grow so that whatever gap that legal scholars had seen in the 1990s between two-earner couples who were white, rather than two-earner couples who were Black, would shrink over time so there would be less of a racial disparity between the two groups. 

Beginning in 2001, Congress took a number of steps that reduced marriage penalties. They doubled the standard deduction and increased rate brackets for married couples relative to unmarried couples. And in 2017, under the Tax Cuts and Jobs Act, more and more married couples were getting that kind of relief. There are still lots of provisions in the tax code that give rise to marriage penalties. But I thought that the combination of more two-earner couples and that changes in the tax code to reduce marriage penalties would mitigate this observation in the ’90s that there was greater occurrence of marriage penalties among Black couples than white couples. 

We looked at marriage penalties over time, prior and post the 2000 and 2017 tax law changes. And we did find that both white and Black couples were less likely to have marriage penalties over time than before 2001. 

But there’s still a gap. 

So the biggest surprise, to me, was that after all the legislation of the past 20 years … under the 2017 tax law, we still observed that Black couples are more likely to incur marriage penalties than white couples. And that the magnitude [of the gap] is a greater percent of income for Black couples than for white couples. 

Q: What do you think still needs to be done to make the tax code more equitable for all couples?

We are beginning to do research where we see more and more inequities in the way in which the tax code affects white taxpayers versus Black taxpayers. That’s an area of research that has not had much attention, in part because we [as a country] think that the tax code is race neutral and we don’t have tax data that distinguishes taxpayers by race. 

All of that is changing now. But because it’s very unlikely, and perhaps undesirable, to have taxpayers indicate their race on their tax return, it’s also unlikely to [have] tax laws that achieve equity between Black and white taxpayers by explicitly referring to race. 

So what are the characteristics that can be observed in the tax code that affect these racial disparities? There are several factors that differentiate Black couples from white couples that contribute to marriage penalties. And some have fixes and some may represent a trade off with more desirable policy goals. 

So one way to fix the marriage penalties, which will affect both Black and white couples, is to allow couples the option of filing as individuals. That would eliminate marriage penalties and it would be very expensive. It might be difficult to allocate which spouse gets unearned income, interest income, dividend income, capital gains, and mortgage interest deductions, or who gets to claim the kids as dependents. It could still open the door for tax avoidance where taxpayers come up with strategies in order to minimize their tax liability in various ways. 

It also cuts away the progressivity of the tax code. Because right now the tax rates go up based on the combined taxpayers’ income. But if you were to tax each individual, you could then end up with a lower tax liability because one of the individuals might get taxed at a low salary and not fully adjust for the resources that the family has in combination. 

The other approach was one done in the 1980s that lasted for about three years, … called the two-earner deduction. It was targeted to a group of people who were most likely to move up into a higher tax rate as a consequence of their combined income. It gave them the opportunity to reduce their income by a portion of the lower earner’s income to alleviate marriage penalties. 

That approach is desirable in that it’s very well targeted. It’s undesirable in that the benefit of the 1980s law was regressive, because the more income [the couple had], their tax rate and the value of that deduction increased. So it was much more valuable to higher-income taxpayers than lower-income taxpayers. 

Another reason we observed marriage penalties more likely among Black couples than white couples is that Black couples are more likely to have dependents. And the reason that matters is that Congress recognizes that families with dependents have greater expenses that make it more difficult for them to pay taxes on the same amount of income as a taxpayer who doesn’t have kids. And so the tax code provides the child tax credit, the earned income tax credit and the head of household filing status. But those provisions actually lead to an increase in marriage penalties.

The head of household filing status gives taxpayers a standard deduction that’s halfway between being single and filing jointly. And that means that if an unmarried single person marries a single parent, the single parent loses a head of household filing status and it makes it more likely that they’re going to have a marriage penalty. 

When a couple gets married, their combined income may result in a smaller or no Earned Income Tax Credit, relative to what they would have gotten filing as unmarried individuals. 

Q: You mentioned that it’s not ideal to eliminate marriage penalties altogether. Why is that? 

Policymakers try to achieve a number of different goals through the tax system. One is tax progressivity, where people pay a higher proportion of their income in taxes as their income goes up. A second goal may be treating families who are similar alike … and [another is] trying to reduce complexity. 

[Marriage] neutrality may be a fourth goal, but tax experts have not found an equation that works. It’s not possible to achieve progressivity, treating similar families alike, and marriage neutrality [with neither penalties nor bonuses]. So policymakers are forced to make trade offs and sometimes those other goals have taken precedence over [eliminating] marriage penalties. In addition, there may be other goals such as tax subsidies. 

For over 100 years, in one way or another, Congress has dealt with how married couples are treated relative to unmarried couples. Sometimes it’s been to the benefit of married couples and sometimes it’s been to the benefit of unmarried couples. 

So this has been a struggle since 1913, when the tax code was first in place. And they still haven’t come up with a solution that achieves all of these goals simultaneously. 

Q: Now that all federal agencies are required to assess how their programs impact racial and ethnic equity under Biden’s Racial Justice Executive Order, do you find that there’s greater access to data around the racial implications of taxes? 

There’s definitely greater attention to how we’re going to measure this. 

Taxpayers are not asked to declare their race on their tax returns, and at least from my perspective, it’s a good thing … [because of]  concerns that it might have dangerous effects on tax enforcement. But on the other hand, without that information, we have been stymied in the past in our efforts to understand how the tax system affects Black and white taxpayers differently. 

So what is happening right now across the government agencies that evaluate tax policy — Department of Treasury, the Joint Committee on Taxation and Congress, the Congressional Budget Office, the Government Accountability Office — is [that they] are developing methodologies to impute race into the tax data. 

And they’re not alone. The Tax Policy Center is also working diligently to come up with methodologies to better understand the racial implications of tax policy. Our study on marriage penalties is an example of one way of doing it, where we started with the [Federal Reserve’s Survey of Consumer Finances] household survey data, created tax units and ran a tax calculator against that. 

But what’s [a bigger] effort for various government agencies and us, is beginning with tax return data, which can give a fuller picture of what’s going on in the tax system, and developing imputations for race and ethnicity to add to our tax analysis. Then we can get this broader picture of the different disparities in a tax code by race and ethnicity. 

Some preliminary results are beginning to emerge, but there’s still a lot of work to be done before we’re able to produce more comprehensive studies of the impact by race and ethnicity. It’s a very important question and a lot of very smart people are working as quickly as they can, but they’re essentially building new infrastructure from scratch.

Help support this work

Public Integrity doesn’t have paywalls and doesn’t accept advertising so that our investigative reporting can have the widest possible impact on addressing inequality in the U.S. Our work is possible thanks to support from people like you. 

Melissa Hellmann is an award-winning reporter who covers racial, gender and economic inequality. Prior...